Part II

The claim that code is law sounds novel. The arrangement behind it is older, and less pure.

Long-distance commerce learned to cross political boundaries before any sovereign possessed the means to govern the whole route. Merchants carried documents, reputations, bonds, and claims among fairs and ports whose courts knew the trade better than distant royal tribunals did. Some forums moved quickly enough to decide a dispute before the market dispersed, and some sanctions traveled far enough to make default costly among people who shared no territorial ruler.

The cleaner legend says that merchants built a universal law of their own and enforced it through reputation alone. The surviving institutions resist the legend. Fair courts drew force from lords, towns, charters, bailiffs, safe conducts, attachable goods, and the possibility of public compulsion. Commercial actors supplied practices, expertise, information, and procedures that public authority did not possess. Private and public power were not rival substances. They were differently placed parts of the same order.

That division gives Part II its first distinction. A court can identify rules, examine evidence, and settle a dispute without possessing a throne, while the coercive backstop may remain elsewhere. Rule-writing and adjudication can detach from territorial monopoly even when enforcement never wholly does. The question is not whether commerce can live without any sword, but where the sword is kept, what claims it will enforce, and whether the forum above it can be disciplined by both sides.

History supplies no founding myth. Private forums can be fast, expert, and portable. They can also become captive to the repeat player who selects them, preserve procedural error with great efficiency, or exclude those who never received standing in the commercial community. Their value lies in the institutional relations they expose: evidence that travels, expertise that remains answerable, sanctions bounded to a domain, exit that does not erase the party who uses it, and a backstop strong enough to make judgment effective without absorbing the judgment into itself.

The firm poses the same problem from another direction. Coase showed that the boundary between hierarchy and market moves with the relative cost of coordinating inside an organization and bargaining outside it. Computational agents lower some of those costs sharply. They can search a larger field, negotiate inside a shared vocabulary, observe performance continuously, and settle through assets already placed within reach. A bounded activity may leave the firm without becoming ungoverned.

The unpaid costs do not disappear. They return in specification, evidence, semantic mismatch, adversarial adaptation, exception, attribution, and persistence. The runtime that accepted a term may vanish while the warranty, loss, or obligation matures later. Cheap execution therefore makes a new constitutional demand visible: some party, fund, or institution must remain answerable after the process ends.

Part II moves from the court that travels to the party that remains. Its subject is not decentralization as a virtue, but the division of institutional work when knowledge, enforcement, and responsibility no longer reside in one place. Computation changes the speed, cost, and reach of that division. It does not abolish power. It moves the places where power must answer.

The next question is political rather than economic. An order may be efficient, plural, and privately constituted while leaving one party dependent upon the continuing permission of another. What distinguishes a useful arrangement from a condition of domination?

Part III begins where permission is tested by adverse will.

---

When enforcement moves, power moves with it.